Deanonimization on the World Wide Web - Closer and Closer
Although not yet so close to start worrying seriously.
At the 29C3 conference in Berlin, Alfa-version of JStylo, a software product that can establish authorship of the text with a rather high probability, was presented by graduate students of Drexel University (USA). It has long been no secret to anyone that a person uses a rather limited set of words for everyday communication, and a profession imposes a certain imprint on this set. The program analyzes messages in social networks, emails, forum posts and tweets, and “tries on” them to the text, the authorship of which must be established. Special attention is paid to words that are used not to describe objects, but to express a connection between them.
While the functionality of the program is rather limited: only English is supported, the circle of “suspects” can be limited to only fifty people, while 6.5 thousand words should be collected on each of them, and the length of the text whose authorship is to be established is not less 500 words (based on NewYorkTimes ).
The same guys are working on another tool - Anonymouth, which should allow to “encrypt” messages, introducing uncharacteristic speech turns into them.
In the future, when the functionality of the program can bring real benefits, it is planned to be used to identify the authors of posts on various thematic hacker forums.
At the 29C3 conference in Berlin, Alfa-version of JStylo, a software product that can establish authorship of the text with a rather high probability, was presented by graduate students of Drexel University (USA). It has long been no secret to anyone that a person uses a rather limited set of words for everyday communication, and a profession imposes a certain imprint on this set. The program analyzes messages in social networks, emails, forum posts and tweets, and “tries on” them to the text, the authorship of which must be established. Special attention is paid to words that are used not to describe objects, but to express a connection between them.
While the functionality of the program is rather limited: only English is supported, the circle of “suspects” can be limited to only fifty people, while 6.5 thousand words should be collected on each of them, and the length of the text whose authorship is to be established is not less 500 words (based on NewYorkTimes ).
The same guys are working on another tool - Anonymouth, which should allow to “encrypt” messages, introducing uncharacteristic speech turns into them.
In the future, when the functionality of the program can bring real benefits, it is planned to be used to identify the authors of posts on various thematic hacker forums.
')
Source: https://habr.com/ru/post/165435/
All Articles